I've been noticing how little I care for categorical statements. So many things are aggregated for no reason at all, it seems. Was Thomas Jefferson a good person? Well, his thoughts on liberty and how to make a country were extremely good – his raping of slaves was extremely bad. How could you possibly compare those things? There is no scale that can weigh them against each other. We know nothing about his quality as a human, what happened when he reached The Pearly Gates – and that's fine. That's not our job.

Why are we even attempting to evaluate an aggregation of a dead person's life? Is it voyeurism? Because people can't be role models unless you can go get groceries with them (my new rule – you need real insight into their lives to know how to fill the role they're modelling). And I'm not even sure they had something we'd recognize as a grocery store back then. You can't learn how to live your life from Jefferson – you'd be impossible hungry without those groceries.

I feel almost all categorical statements should be replaced by specific statements about the issue at hand. But there are some dangers: what do we do about less-than-wise decisions, and what do we do about principles? How do we learn about things and how do we imagine them?

First off, it's worth saying we should be more comfortable defending our principles in action. If you think people should be able to say whatever they want, step in front of them when the government wants to punish them for making a dog joke (unlike the celebrities that will speak out in support in cases like this, but do nothing). If you think people should have reasonable availability of guns, defend a responsible gun owner killed for no reason other than their possession of a firearm (unlike the NRA, who did not defend the gun owner).

Defend principles in the particulars, because we learn what it means to have principles when our heart tells us even the extreme cases are worth defending. That's what it means to be principled. And, while it's pretty rare to have that as an aesthetic, it's a complicated world, and the marginal value of solid principles is only rising. Do you think you can predict all the consequences of a complex law? Then stop trying; try Principles instead.

A complex world is a mine-field for careful thinkers, weighing things in the balance, and trying to make the wisest, most nuanced decision. There's simply too much information! It may well be impossible for technocratic decision-making to achieve good results. This arguably started over 50 years ago – but who knows? – and computers certainly didn't make it easier.

When we develop our principles, we need to imagine them in some way. What is our emblem for guns? Is it a mass shooting (extremely unlikely, and n.b. extremely ill-advised), or is it defending against a potential tyranny (extremely unlikely, and n.b. ill-advised if you're using modern civilian weapons against a terrifyingly well armed military)? Or are those exceptions? What's the median use? How frequently are guns (or the threat of gun use) used defensively, and how many people use them to commit crimes? How many terrified and traumatized people get substantial mental health benefits from finally feeling safe, vs. how many use it as a way to act on a momentary suicidal desire?

I don't have the answers – I barely have a decent handful of questions. But people are very eager to change broad-use policies, make them complex, but never dive into details. Should this person be put at odds with the law? It's the only question their approach can answer, but we do a disservice to assume it can be answered perfectly. Make a systen where you can live with making some mistakes, because you'll make them. Try to avoid systems where it will punish people without reason or recourse – because our principles of fairness still apply. Type I and Type II errors are inevitable, but each person gets their own disaggregated defence of fairness.